At Sovereign Intelligence, we are asked on a regular basis about the effects of AI and machine learning on Policing often.
In response, we took a deeper dive into this question and after reviewing our stance on AI risk in law enforcement, the team at Sovereign identified five leading risks that policing could face if AI technology was effectively used by malicious actors: Cyberattacks, Misinformation, Autonomous Weapons, Detection Evasion, and Counterfeiting.
THE RISKS, DESCRIBED
Cyberattacks (Probability: FREQUENT / Severity: CATASTROPHIC) are already an attack vector for malicious actors. These attacks exploit vulnerabilities in law enforcement systems and have historically been assisted by a vibrant human community. With the addition of smart AI tooling purpose built for cyberattacks to the community’s toolset, the risk of a breach exponentially increases. Breaches could result in access to sensitive data, disruptions in investigations/operations, and potential risks to officer safety.
Manipulation and misinformation (Probability: LIKELY / Severity: CATASTROPHIC), is yet another well-known set of techniques that can be turbo-charged with AI-assisted capabilities. Coupled with social media and facilitated by “deepfake” technology, fabricated video/pictorial evidence or distorted narratives could undermine investigations and erode trust in legal proceedings. The social aspect of this risk could lead to catastrophic outcomes for civil order and higher-order government processes as a whole. These techniques have already been demonstrated against citizens for extortion, bribery, and blackmail purposes. Levied against law enforcement, we expect a disruptive outcome.
While we consider this to be an occasional issue and only applicable to certain operational circumstances, AI-powered autonomous weapons through the use of weaponized drones (Probability: OCCASIONAL / Severity: CATASTROPHIC) have become very easy to fabricate and operate which raises concerns of criminal misuse. AI-capable drones equipped with explosives or ballistics, pose a threat to both public critical infrastructure and police operations.
Smarter malicious actors will likely understand law enforcement TTPs to a degree that they can use AI-enhanced tools for detection and surveillance evasion (Probability: LIKELY / Severity: MARGINAL). We only consider this to be a marginal severity due to the number of other TTPs available to law enforcement as a fall-back measure. This does not mean that law enforcement should ignore the increasing risk of detection evasion.
Like manipulation and misinformation, AI-generated counterfeit documents (Probability: LIKELY / Severity: MARGINAL) and digital signatures pose risks of increased incidents of forgery and fraud. While this does not have the probability of a direct effect on law enforcement operations like the previously listed risks, criminals can exploit these tools for more effective and higher-volume identity theft and money laundering operations. We expect the outcome to be in higher incidence rates and increased caseloads in this area.
PREPARING FOR AN AI FUTURE IN LAW ENFORCEMENT
Per our prioritized risks above, law enforcement should consider ramping up technological investment in the following areas:
Cybersecurity Tools & Threat Intelligence: Bolstering policing system defenses against hacking, tampering, and unauthorized access while using tools that create situational awareness of looming threats before they materialize.
Deepfake & Misinformation Identification Tooling: Develop AI tools or procure COTS solutions for law enforcement to help them identify deepfakes, misinformation, and fake documents.
Anti-Drone and Automated Countermeasures: Develop or procure COTS anti-drone capabilities for law enforcement
Law enforcement should also consider TTP/Program investment in the following areas:
Traditional Investigation/Policing Techniques: With the advent of new tools for policing, current operations and culture may change based on increased efficiencies and capabilities. Ensure that there is still a strong culture of “traditional policing” in every AI-enabled operation.
Education & Collaboration: Collaboration is a pillar of effective policing. Extending collaboration to academia, the private sector, and other agencies allows law enforcement to collectively learn about and defend against AI-driven crimes. Sharing insights and practices fosters innovation and prepares policing units for unknown unknowns. In a borderless digital world, global collaboration is vital. Partnering with international law enforcement agencies creates a united approach against cross-border AI-driven crimes, addressing challenges together.
The future of policing and law enforcement in general can benefit from AI Technologies as other industries continue to benefit. As law enforcement agencies begin to adopt defences against bad actors and the technology for use in policing, we expect the conversation about AI to evolve. Contact us at any time to start a discussion about how Sovereign Intelligence can help you along the way to this inevitable future at: firstname.lastname@example.org